Validation of Multiple Imputation in SPHC-B 2010

Notes: This document summarizes only the key outputs of the final multiple imputation process for SPHC-B 2010. More detailed R scripts are available in the R Notebook “SPHC-B-2010.Rmd”.

1. Load Packages

library(naniar)
library(dplyr)
library(jomo)
library(mitml)
library(ggplot2)

2. SPHC-B 2010

2.1. Incomplete dataset

load("/Volumes/LGBT Project data/Multiple Imputation/d_2010_incomplete.RData")
summary( d_2010_incomplete )

    sampling_strata_region calibrated_weight no.of.population   sexual_identity_2010 sexual_identity_fluidity
 Lidingö       : 1368      Min.   :  5.424   Min.   :  6831   Heterosexual:28471     0   :11534              
 Nacka         : 1342      1st Qu.: 25.180   1st Qu.: 26849   Homosexual  :  361     1   : 1656              
 Nynäshamn     : 1284      Median : 43.971   Median : 34342   Bisexual    :  381     NA's:17577              
 Upplands-Väsby: 1268      Mean   : 52.157   Mean   : 41143   Uncertain   :  394                             
 Sundbyberg    : 1144      3rd Qu.: 70.027   3rd Qu.: 56341   NA's        : 1160                             
 Botkyrka      : 1082      Max.   :282.667   Max.   :101916                                                  
 (Other)       :23279                                                                                        
      age             sex              country_of_birth       education         income                 marital_status 
 Min.   : 18.00   Male  :13829   Sweden        :25157   <=9 years  : 4998   Min.   :     0   Never married    : 9887  
 1st Qu.: 37.00   Female:16938   Europe        : 3343   10-12 years:12548   1st Qu.:  1917   Currently married:15066  
 Median : 51.00                  Outside Europe: 2267   >=13 years :13001   Median :  2696   Other            : 5814  
 Mean   : 51.04                                         NA's       :  220   Mean   :  3247                            
 3rd Qu.: 65.00                                                             3rd Qu.:  3667                            
 Max.   :104.00                                                             Max.   :450719                            
                                                                                                                      
 living_alone personal_support weight_strata  
 yes : 6432   yes :26882       2      : 1238  
 no  :24005   no  : 3474       16     : 1235  
 NA's:  330   NA's:  411       8      : 1232  
                               11     : 1232  
                               14     : 1232  
                               18     : 1232  
                               (Other):23366

sapply( d_2010_incomplete, class ) # all continuous variables are numeric, and all categorical variables are factor

  sampling_strata_region        calibrated_weight         no.of.population     sexual_identity_2010 
                "factor"                "numeric"                "numeric"                 "factor" 
sexual_identity_fluidity                      age                      sex         country_of_birth 
                "factor"                "numeric"                 "factor"                 "factor" 
               education                   income           marital_status             living_alone 
                "factor"                "numeric"                 "factor"                 "factor" 
        personal_support            weight_strata 
                "factor"                 "factor"

miss_var_summary( d_2010_incomplete ) # 57.1% missing in sexual_identity_fluidity, 3.8% in sexual_identity_2010, 1.3% in personal_support, 1.1% in living_alone, and 0.7% in education

2.2. Two-level multivariate normal imputation

# specify imputation model
# fml_imp_2010 <- sexual_identity_fluidity + sexual_identity_2010 + personal_support + living_alone + education ~ 1 + age*sex + country_of_birth + income + marital_status + ( 1 | weight_strata )

# final imputation with the chosen number of iterations
# imp_final_2010 <- jomoImpute( data = d_2010_incomplete,
#                               formula = fml_imp_2010,
#                               random.L1 = "full",
#                               n.burn = 2000,
#                               n.iter = 1000,
#                               m = 80,
#                               seed = 12345
#                               ) # took around 20 hours

load("/Volumes/LGBT Project data/Multiple Imputation/imp_final_2010.RData")
summary( imp_final_2010 ) # summarize model and display convergence statistics


Call:

jomoImpute(data = d_2010_incomplete, formula = fml_imp_2010, 
    random.L1 = "full", n.burn = 2000, n.iter = 1000, m = 80, 
    seed = 12345)

Cluster variable:         weight_strata 
Target variables:         sexual_identity_fluidity sexual_identity_2010 personal_support living_alone education 
Fixed effect predictors:  (Intercept) age sex country_of_birth income marital_status age:sex 
Random effect predictors: (Intercept) 

Performed 2000 burn-in iterations, and generated 80 imputed data sets,
each 1000 iterations apart. 

Potential scale reduction (Rhat, imputation phase):
 
         Min   25%  Mean Median   75%   Max
Beta:  1.000 1.000 1.004  1.001 1.004 1.028
Psi:   1.000 1.000 1.000  1.000 1.000 1.001
Sigma: 1.000 1.000 1.073  1.012 1.056 1.989

Largest potential scale reduction:
Beta: [1,2], Psi: [2,1], Sigma: [42,1]

Missing data per variable:
    weight_strata sexual_identity_fluidity sexual_identity_2010 personal_support living_alone education
MD% 0             57.1                     3.8                  1.3              1.1          0.7      
    sampling_strata_region calibrated_weight no.of.population age sex country_of_birth income marital_status
MD% 0                      0                 0                0   0   0                0      0

plot( imp_final_2010, trace = "all", print = "beta" ) # check trace and autocorrelation plots

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

2.3. Validate imputed data

# extract imputed datasets
original_data_2010 <- mitmlComplete( imp_final_2010, print = 0 ) # extract original incomplete dataset
implist_2010 <- mitmlComplete( imp_final_2010, print = "all" ) # extract all imputed datasets

original_data_2010$imputation <- "0"
all_data_2010 <- bind_rows( original_data_2010,
                            bind_rows( implist_2010, .id = "imputation" ) ) # merge datasets
all_data_2010$imputation <- as.numeric( all_data_2010$imputation )
summary( all_data_2010 )

 weight_strata     sexual_identity_fluidity   sexual_identity_2010 personal_support living_alone         education      
 2      : 100278   0   :2112580             Heterosexual:2391929   yes :2205638     yes : 529530   <=9 years  : 409994  
 16     : 100035   1   : 361970             Homosexual  :  30755   no  : 286078     no  :1962267   10-12 years:1023200  
 8      :  99792   NA's:  17577             Bisexual    :  32530   NA's:    411     NA's:    330   >=13 years :1058713  
 11     :  99792                            Uncertain   :  35753                                   NA's       :    220  
 14     :  99792                            NA's        :   1160                                                        
 18     :  99792                                                                                                        
 (Other):1892646                                                                                                        
    sampling_strata_region calibrated_weight no.of.population      age             sex                country_of_birth  
 Lidingö       : 110808    Min.   :  5.424   Min.   :  6831   Min.   : 18.00   Male  :1120149   Sweden        :2037717  
 Nacka         : 108702    1st Qu.: 25.180   1st Qu.: 26849   1st Qu.: 37.00   Female:1371978   Europe        : 270783  
 Nynäshamn     : 104004    Median : 43.971   Median : 34342   Median : 51.00                    Outside Europe: 183627  
 Upplands-Väsby: 102708    Mean   : 52.157   Mean   : 41143   Mean   : 51.04                                            
 Sundbyberg    :  92664    3rd Qu.: 70.027   3rd Qu.: 56341   3rd Qu.: 65.00                                            
 Botkyrka      :  87642    Max.   :282.667   Max.   :101916   Max.   :104.00                                            
 (Other)       :1885599                                                                                                 
     income                 marital_status      imputation
 Min.   :     0   Never married    : 800847   Min.   : 0  
 1st Qu.:  1917   Currently married:1220346   1st Qu.:20  
 Median :  2696   Other            : 470934   Median :40  
 Mean   :  3247                               Mean   :40  
 3rd Qu.:  3667                               3rd Qu.:60  
 Max.   :450719                               Max.   :80

# sexual identity in 2010
ggplot( all_data_2010[ !is.na( all_data_2010$sexual_identity_2010 ), ],
        aes( fill = sexual_identity_2010, x = imputation ) ) + 
  geom_bar( position = "fill" ) + 
  scale_y_continuous( labels = scales::percent ) + 
  scale_fill_discrete( name = "Sexual identity in 2010" ) +
  labs(
    x = "Imputation number",
    y = "Proportion",
    caption = "Notes: Imputation number 0 represents the original incomplete dataset." ) +
  theme_classic() +
  theme( axis.title.x = element_text( family = "Arial", size = 11 ),
         axis.text.x = element_text( family = "Arial", size = 11 ),
         axis.text.y = element_text( family = "Arial", size = 11 ),
         axis.title.y = element_text( family = "Arial", size = 11 ),
         legend.text = element_text( family = "Arial", size = 10 ),
         legend.title = element_text( family = "Arial", size = 10 ),
         legend.position = "bottom",
         plot.caption = element_text( family = "Arial", size = 10, hjust = 0 ) 
  )


# change in sexual identity during 2010-2021
ggplot( all_data_2010[ !is.na( all_data_2010$sexual_identity_fluidity ), ],
        aes( fill = sexual_identity_fluidity, x = imputation ) ) + 
  geom_bar( position = "fill" ) + 
  scale_y_continuous( labels = scales::percent ) + 
  scale_fill_discrete( name = "Change in sexual identity during 2010-2021", labels = c( "No", "Yes" ) ) +
  labs(
    x = "Imputation number",
    y = "Proportion",
    caption = "Notes: Imputation number 0 represents the original incomplete dataset." ) +
  theme_classic() +
  theme( axis.title.x = element_text( family = "Arial", size = 11 ),
         axis.text.x = element_text( family = "Arial", size = 11 ),
         axis.text.y = element_text( family = "Arial", size = 11 ),
         axis.title.y = element_text( family = "Arial", size = 11 ),
         legend.text = element_text( family = "Arial", size = 10 ),
         legend.title = element_text( family = "Arial", size = 10 ),
         legend.position = "bottom",
         plot.caption = element_text( family = "Arial", size = 10, hjust = 0 ) 
  )


# education
ggplot( all_data_2010[ !is.na( all_data_2010$education ), ],
        aes( fill = education, x = imputation ) ) + 
  geom_bar( position = "fill" ) + 
  scale_y_continuous( labels = scales::percent ) + 
  scale_fill_discrete( name = "Level of education" ) +
  labs(
    x = "Imputation number",
    y = "Proportion",
    caption = "Notes: Imputation number 0 represents the original incomplete dataset." ) +
  theme_classic() +
  theme( axis.title.x = element_text( family = "Arial", size = 11 ),
         axis.text.x = element_text( family = "Arial", size = 11 ),
         axis.text.y = element_text( family = "Arial", size = 11 ),
         axis.title.y = element_text( family = "Arial", size = 11 ),
         legend.text = element_text( family = "Arial", size = 10 ),
         legend.title = element_text( family = "Arial", size = 10 ),
         legend.position = "bottom",
         plot.caption = element_text( family = "Arial", size = 10, hjust = 0 ) 
  )


# living status
ggplot( all_data_2010[ !is.na( all_data_2010$living_alone ), ],
        aes( fill = living_alone, x = imputation ) ) + 
  geom_bar( position = "fill" ) + 
  scale_y_continuous( labels = scales::percent ) + 
  scale_fill_discrete( name = "Living alone", labels = c( "Yes", "No" ) ) +
  labs(
    x = "Imputation number",
    y = "Proportion",
    caption = "Notes: Imputation number 0 represents the original incomplete dataset." ) +
  theme_classic() +
  theme( axis.title.x = element_text( family = "Arial", size = 11 ),
         axis.text.x = element_text( family = "Arial", size = 11 ),
         axis.text.y = element_text( family = "Arial", size = 11 ),
         axis.title.y = element_text( family = "Arial", size = 11 ),
         legend.text = element_text( family = "Arial", size = 10 ),
         legend.title = element_text( family = "Arial", size = 10 ),
         legend.position = "bottom",
         plot.caption = element_text( family = "Arial", size = 10, hjust = 0 ) 
  )


# personal support
ggplot( all_data_2010[ !is.na( all_data_2010$personal_support ), ],
        aes( fill = personal_support, x = imputation ) ) + 
  geom_bar( position = "fill" ) + 
  scale_y_continuous( labels = scales::percent ) + 
  scale_fill_discrete( name = "Personal support", labels = c( "Yes", "No" ) ) +
  labs(
    x = "Imputation number",
    y = "Proportion",
    caption = "Notes: Imputation number 0 represents the original incomplete dataset." ) +
  theme_classic() +
  theme( axis.title.x = element_text( family = "Arial", size = 11 ),
         axis.text.x = element_text( family = "Arial", size = 11 ),
         axis.text.y = element_text( family = "Arial", size = 11 ),
         axis.title.y = element_text( family = "Arial", size = 11 ),
         legend.text = element_text( family = "Arial", size = 10 ),
         legend.title = element_text( family = "Arial", size = 10 ),
         legend.position = "bottom",
         plot.caption = element_text( family = "Arial", size = 10, hjust = 0 ) 
  )

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